What's Changed
This release brings two major tracing capabilities, broader export options, and a round of security and stability work.
Features
Multi-pod tracing & cross-namespace support (#63)
Trace workloads that span multiple pods across namespaces in a single session.
Performance profiling integration (#67)
Slow events now come with the exact goroutine and CPU stacks that caused them. Podtrace fetches pprof data (heap, goroutine, CPU) from target pods and correlates it with BPF stack traces, auto-triggered on configurable latency thresholds, or on demand via /profile/start`. BPF kernel timestamps are now aligned to wall-clock for precise correlation with application logs.
New trace exporters: Datadog and Zipkin (#68)
In addition to OTLP, Jaeger, and Splunk HEC, you can now export traces directly to Datadog and Zipkin backends.
Security & Stability (#66)
- Hardened Podtrace security posture
- Updated dependencies
🧪 Quality & Tooling (#67, #68)
- Expanded test coverage
- Upgraded CI workflows
- Consolidated documentation into a single source
Upgrading
No breaking changes.
Full Changelog: v0.9.0...v0.10.0